Eficiencia Computacional

The understanding of organizational knowledge as a source of operational

performance and sustainability has increased in the public sector (Jain & Jeppesen, 2013).

Leaders must focus on creating and enhancing knowledge-sharing processes (Chong et al., 2011; Pinho et al., 2012). The objective of this study was to examine the nature and extent of the relationship between (a) employee trust, organizational fairness, and supervisor competency and (b) employee’s knowledge sharing. In this section, I include detailed information on the methodology and research process, (a) Purpose Sstatement, (b) Role of the Researcher, (c) Participants, (d) Research Method and Design, (e) Population and Sampling, (f) Data Collection, (g) Reliability and Valadity, and (h) Summary.

Purpose Statement

The purpose of this quantitative correlational study was to identify the extent and nature of the correlation between (a) employee trust, organizational fairness, and

supervisor competency and (b) the willingness of employees in public housing authorities in Texas to share knowledge. Considering that people are a vital element of the

knowledge-sharing process, leaders need to examine the culture of the organization to learn how much it has a supportive and effective knowledge-sharing environment

(Deverell & Burnett, 2012). The analysis of standard multiple regression and significance of correlation of the independent variables on knowledge-sharing willingness may assist leaders in promoting knowledge friendly working environments. Leaders of public housing authority agencies might utilize the study findings to establish effective

knowledge sharing processes. Effective knowledge sharing processes assists leaders in collecting organizational wisdom and can contribute to intellectual capital retention amongst employees (Turner et al., 2012). Resultant improvements in performance at public housing authorities could (a) expand the housing service to low-income residents, (b) reduce taxpayers’ burden by effectively improving business processes, and (c)

increase social service quality by enforcing the compliance of HUD’s sustainability plan.

Role of the Researcher

As the researcher, I actively involved myself in all processes of this study, including (a) data collection, (b) storage, (c) analysis, (d) data integrity, (e) confidentiality, and (f) the proffer of conclusions. I reviewed the Belmont Report

protocol and completed Protecting Human Research Participants training by the National Institutes of Health (NIH) Office of Extramural Research (certification number 803591).

The study components included the development and verification of the survey questions, performing the pilot study, and conducting the final study.

Interaction between social actions sustains knowledge (Pillay & James, 2014).

From the constructivism worldview, practitioners focus on active participants by conducting and communicating knowledge creation in an organization amongst employees (Yoo, Kim, & Kwon, 2014). Ensuring a freedom to participate in this quantitative study, I did not influence the population with knowledge and experience regarding the housing authority business.

I serve as the Director of Information Technology Resources for a local housing authority where I have implemented available technologies to improve the agency’s

business operations and procedural processes. I am familiar with HUD’s regulatory requirements and sustainability policy. In this position, I have a professional relationship with leaders and employees in my agency. However, I do not have any relationship with the employees and leaders in other agencies in the State of Texas. Bias causes a

misrepresentation of the result findings and can occur in any assessment of data

collection process (Healy & Devane, 2011). Becker (2013) stated that avoiding contact with participants prior to the survey ensures preconception do not occur. To manage potential bias, I did not include my agency in the study population.

Participants

Employee motivation and collaboration will positively affect knowledge sharing in an organization (Rasula et al., 2012). However, there is a gap in the literature focusing on knowledge sharing in the public sector (Amayah, 2013). The target population for this study consisted of fulltime employees and leaders in public housing authorities in the State of Texas. I used purposive sampling to assure the participants’ relevance to the research questions (Bryman, 2012). As a Director of Information Technology Resources for a local housing authority, I understand how knowledge-sharing processes occur in organizations. This understanding assisted in building relationships with employees and executive leaders across housing authorities in the State of Texas.

After IRB approval (No. 10-13-14-0250051), I sent an introduction letter

(Appendix B) to all executive directors of public housing authority agencies in the State of Texas asking permission to conduct a survey among their employees and leaders.

Public housing authority listings are publically available through the HUD website. Once

executive directors agreed to participate in this study, I requested that an authorized representative of each housing authority send the online survey link to the target

population. The online survey host was Survey Monkey®. Because data analyses based on individual local housing authorities did not occur, leaders participated without employees in the same location and employees participated without respective leaders.

Since an authorized representative distributed the link to the online survey, there was no identifiable information requirement; however housing authority executive directors, authorized representatives, and city/county demographics remained confidential. All participants completed a consent form to participate. Participants could withdraw from the study at any time, and until final response submission. I have sole access to all data, saved in an USB drive and stored in a locked, fireproof safe for a period of 5 years.

Research Method and Design

For this study, I used a quantitative correlational design to examine the

relationship between employee trust, organization fairness, and supervisor competency on knowledge-sharing behaviors. Muijs (2011) suggested that researchers whose worldview underlies positivism, experiential realism, or pragmatism tend to use a quantitative methodology in natural or social science studies. Quantitative research is an investigative tool that researchers use to examine descriptions of phenomena, changes over time within groups, or relationships amongst variables including predictions (Rovai, Baker, & Ponton, 2013). Experimental and nonexperimental are two types of quantitative research designs used to test or examine the validity of a hypothesis (Muijs, 2011).

According to Rovai et al. (2013), nonexperimental designs include descriptive,

correlational, and causal-comparative designs. In conducting correlational research, investigators can examine relationships between two or more existing and

nonmanipulating variables (Green & Salkind, 2011).

Research Method

Qualitative, quantitative, and mixed methods are different approaches to conducting a research study (Rovai et al., 2013). Applying quantitative methodology, investigators confirm a linkage amongst sets of (a) data, (b) business factors, (c) financial success, or (d) management performance (Malina et al., 2011). Muijs (2011) stated that researchers employ quantitative methods to collect and mathematically analyze data to explain a particular phenomenon. Moreover, quantitative researchers test a theory or hypothesis to explain relationships between independent and dependent variables (Allwood, 2012; Malina et al., 2011). Likewise, Chong et al. (2011) conducted a quantitative study to test the correlation between organizational factors and the willingness to share knowledge in public sector organizations in India. In addition, Husted et al. (2012) used a quantitative research method to examine the relationship between organizational governance and knowledge-sharing behavior. For this study, I used a quantitative method to examine the correlational relationship of employee trust, organizational fairness, and supervisor competency on the willingness to share

knowledge. Therefore, a quantitative method was suitable for this study.

Researchers use qualitative methods to explore perceived meanings, leading to an interpretive estimation of the existing phenomena (Fuhse & Mutzel, 2011) and to

understand social problems (Savage-Austin & Honeycutt, 2011). In addition, qualitative

researchers explore the experiences of research participants rather than a researcher’s topic (Fisher & Stenner, 2011). Rusly et al. (2014) adopted a qualitative methodology to assess the influence of change perceptions on knowledge-sharing processes in the business environment. Since the purpose of this study was to examine relationships instead of perceived meanings, a qualitative method was not appropriate.

Mixed methods researchers blend qualitative and quantitative methods (Muijs, 2011). Researchers use mixed-methods to examine and explore causality and meanings (Muijs, 2011). According to Bryman (2012), researchers use mixed methods when the focus on the phenomenon is an issue of mathematical clarity by comparing qualitative and quantitative findings. Since I only employed numerical analysis, absent of a phenomenon, a mixed-method approach was not suited for this study.

Research Design

Quantitative experimental designs provide researchers with strong claims for causality through the utilization of the ability to assign random value for the factors used to manipulate values of variables (Whitley & Kite, 2013). Conversely, quantitative non-experimental designs are suited for investigating relationships between variables occurring in a particular context (Muijs, 2011). Since the purpose of the study was to examine linear correlations of employee trust, organizational fairness, and supervisor competence on the willingness to share knowledge amongst employees, a quantitative non-experimental design was appropriate. Because experimental designs are the strongest approach for addressing internal validity, researchers use experimental designs to

determine causality (Whitley & Kite, 2013). Moreover, experimental designs involve

manipulation of variable’s values to find the effects of one variable to another (Field, 2013). Because I could not manipulate the values of the variables in this study, experimental designs were not appropriate.

Nonexperimental designs include descriptive, correlational, and

causal-comparative or ex post facto (Rovai et al., 2013). Researchers use descriptive designs to generate records for a phenomenon within a given population (Muijs, 2011). A

correlational design is appropriate for investigators to examine relationships or prediction between variables (Whitley & Kite, 2013). Pangil and Chan (2014) chose a regression analysis to test the correlations between knowledge-sharing relationships with trust and virtual team effectiveness. Researchers who use causal-comparative design, or ex post facto design, examine possible causes or consequences of differences (Rovai et al., 2013).

I used a correlational design to test hypotheses and to determine the prediction existed between the independent variables and dependent variable.

A correlation design is appropriate to measure variable relationships (Pallant, 2013). In addition, Wallen and Fraenkel (2013) noted that quantitative researchers employ correlational designs to examine essential human behaviors or predict likely outcomes based on variables’ relationships. Carmeli et al. (2013) conducted a regression analysis to examine the relationship between leadership and creativity to mediate the role of knowledge sharing.

Researchers use the statistical significance of the correlation coefficient to calculate the likelihood of a relationship between two studied factors (Bryman, 2012).

Therefore, I conducted a data analysis using a standard multiple regression and

correlation with IBM SPSS^® 22.0 (Pallant, 2013) to study the prediction of multiple variables and to test each of the hypotheses. Although the purpose of this study was to examine a linear relationship between variables, I also conducted a descriptive analysis to understand the demographics of the participants (Green & Salkind, 2011). Additionally, a regression model test for the prediction of knowledge-sharing willingness from employee trust, organizational fairness, and supervisor competency supported the study findings.

Amayah (2013) used a multiple regression analysis to examine the determinants of knowledge sharing in a public sector organization. I analyzed a standard multiple regression model to address two questions relating to the central research question for this study:

 How do the three independent variables of trust, fairness, and competency predict knowledge-sharing behavior?

 Which, if any, is the best predictor of knowledge-sharing behavior: employee

trust, organizational fairness, or supervisor competency?

Population and Sampling

Public housing authority agencies vary in sizes, scopes, and organizational structure (Kumar & Bauer, 2010). According to HUD (2014), 413 housing agencies represent many local cities and towns in the State of Texas. The population consisted of employees and leaders employed fulltime by public housing authority agencies in the State of Texas. Researchers use purposive sampling to ensure the credibility of potential participants (Becker, 2013). Purposive sampling allows the researcher to collect rich data and increase study validity (Suri, 2011). Moreover, Hoch (2014) employed purposive

sampling to select quantitative data from 280 team members of a medium sized business development provider to examine the influence of leadership on knowledge sharing. I used a purposive sampling method to identify the target population to examine if a correlational existed between trust, fairness, and competence with knowledge-sharing willingness. I sent an introduction letter (Appendix B) regarding the purpose of the study to all executive directors of public housing authority agencies in the State of Texas requesting permission to conduct a survey of employees and leaders. Public housing authority listings and contact information were publically available through the public HUD website (HUD, 2014). After agreeing to allow their agency to participate in this study, the executive director designated an authorized representative of each authority to send an online survey link via e-mail, along with a brief overview of the research, to the target population. The online survey host was Survey Monkey®. The participants could access the survey from any geographical location.

Since each of the values of employees’ trust, organizational fairness, or supervisor competency was random, I conducted a random effect multiple regression model. All three hypotheses H1a, H2a, and H3a were directional. Field (2013) suggested that researchers conduct a one-tailed statistical test for a directional hypothesis.

In quantitative research, the determination of the sample size is necessary for the interpretation of a correlational strength between variables (Field, 2013). Effect size, alpha value, and statistical power are the parameters for calculating the sample size (Muijs, 2011). The reliability of research findings is dependent on an adequate sample size (Wallen & Fraenkel, 2013). Cohen (1992) analyzed statistical power in research to

provide the effect sizes and sample sizes required for power = .80 to detect the effects via various statistical tests. Effect size index and value for small, medium, and large effect are imperative in determining of population sample size for quantitative analysis (Cohen, 1992). Relating to the prediction in multiple regression testing, Cohen (1992) defined the values for small, medium, and large effect size index respectively as .02, .15, and .35.

Explaining further, Cohen suggested that the actual medium effect size is .1304.

Therefore, the medium effect size .15 used in G*Power software to calculate the sample size was about 13% greater than Cohen’s actual medium effect size of .1304. I employed a power test analysis to calculate the sample size required for the study (Field, 2013) and conduct a power analysis with a linear multiple regression, random effect model (exact F-test). The sample size generated by G*Power 3.1.2 software for conducting 1-tailed test in this study (Faul et al., 2009) where α = .05, power = .80, and effect size = .15 for three predictors was 69 (Appendix C).

Ethical Research

Codes of conduct guidelines are essential for handling and directing research (Muijs, 2011). Ethical research includes (a) informed consent, (b) voluntary participation, (c) harm prevention, (d) confidentiality, and (e) protection of vulnerable populations (Rovai et al., 2013). In addition, Whitley and Kite (2013) categorized ethical research as respect, beneficence, and justice. Respect refers to voluntary participation, informed consent, and freedom to withdraw from participation (Whitley & Kite, 2013).

Beneficence means the protection of vulnerable populations, avoidance of harm, and confidentiality (Whitley & Kite, 2013). Justice also refers to informed consent and

voluntary participation (Whitley & Kite, 2013). Ethical considerations are guidelines for all researchers.

After obtaining an agreement from the participating housing authorities, an authorized representative invited all participants meeting the criteria for the study to complete an online survey via Survey Monkey®. Online survey pages were not available until the participant confirmed the agreement to participate on the first page of the survey link. This confirmation served as implied consent by the participants. Participants could withdraw from the study at any time prior to the final submission of the survey by

refusing to complete or terminating the survey. There were no incentives to participate or requirements for the names of individual employees or respective housing agencies. Any information regarding the name of executive directors who agreed to the study,

authorized representative, or county/city identification remains confidential. I have sole access to all data, saved in an USB drive and stored in a locked, fireproof safe for a period of 5 years.

Data Collection Instruments

In quantitative studies, the Likert scale is a measurement that can assist researchers with the value of variables’ information (Rovai et al., 2013). I used the 5-point Likert questions to gather data responses. Rating scales such as the Likert-type provide respondents the ability to indicate the degree to which they agree with the statement item (Muijs, 2011). In addition, quantitative researchers use Likert-type surveys in establishing equally weighted statements regarding participants’ perception,

attitudes, or opinions (Rovai et al., 2013). The survey question response options were choices among five levels of agreement: strongly disagree, disagree, neutral, agree, and strongly agree. The scores of the responding values respectively ranked from 1 to 5.

The online survey consisted of two parts and a total of 45 questions (Appendix A) and was hosted by Survey Monkey®. Part 1 contained questions to generate anonymous demographic information. To understand the demographics of the population, I

conducted a descriptive analysis. Within quantitative methods, demographic data are required for conducting descriptive analyses (Green & Salkind, 2011). Part 2 included survey questions to obtain responses for the values of predictors and for testing the hypotheses.

To assure the instrument’s validity, I adopted survey instrument based on an extensive review of available peer reviewed literature on the topic. Demonstrating construct validity requires testing of the instrument derived, based on the hypothesis and research questions (Tabachnick & Fidell, 2013). Quantitative investigators explore construct validity by examining the related (convergent validity) and unrelated (discriminant validity) relationship of the constructed variables (Pallant, 2013). To address the concerns with construct validity, convergent validity, and discriminant

validity, I adopted the measurement indicators from peer reviewed literature and obtained permission to reuse the text from the publishers (Appendix D), regarding (a) employees’

willingness to share knowledge, (b) social networks, (c) supervisor competency, and (d) organizational factors. For each of the measurement indicators, I reused 5-point Likert scale survey questions from the previous studies. Table 2 contains a summary of how the

instrument items related to the measurement indicators of the available peer reviewed literature.

Table 2

Survey Instrument Questions Relationship to Literature

Literature sources Measurement indicators Survey questions

Kim and Lee (2010) Social networks ET1, ET2, ET3, ET4, ET5, ET6, ET7, and ET8.

Kim and Lee (2010), Performance based OF9, OF10, OF11, OF12, OF13, Reychav and Sharkie award, reward OF14, and OF15.

(2010) expectation, and

intrinsic job motivation

Byrne et al. (2012) Trust in supervisor SC16, SC17, SC18, SC19, SC20, SC21, SC22, SC23, and SC24.

Byrne et al. (2012), Knowledge-sharing KS25, KS26, KS27, KS28, KS29, Kim and Lee (2010), willingness KS30, KS31, KS32, KS33, KS34,

Reychav and Sharkie KS35, KS36, KS37, KS38, KS39

(2010) and KS40

The purpose of collecting data from survey questions 1 to 8 was to examine the employees’ perception of trust, coding as ET1, ET2, ET3, ET4, ET5, ET6, ET7, and ET8.

Questions’ 9 to 15, coding as OF9, OF10, OF11, OF12, OF13, OF14, and OF15, related to the perception of participants regarding organizational fairness. The responses to questions 16 to 24, coding as SC16, SC17, SC18, SC19, SC20, SC21, SC22, SC23, and SC24, revealed employee perceptions of their supervisor’s competency. Question 25 to 40 measured the degree of the willingness of employees to engage in knowledge sharing, coding as KS25, KS26, KS27, KS28, KS29, KS30, KS31, KS32, KS33, KS34, KS35,

KS36, KS37, KS38, KS39 and KS40. A copy of the instrument is located in Appendix A.

Two design types in descriptive studies are cross-sectional and longitudinal (Rovai et al., 2013). Based upon the nature of this study, my instrument followed the cross-sectional design. The anticipated data collection timeframe for the pilot study was 1week, and the length of data collection process for the final study was 2 weeks.

Table 3

Pilot Study - Cronbach’s Alpha Coefficients for Sets of Questions

Question set N Mean SD Cronbach’s

Before I proceeded with the final study, I performed a test of the instrument for validity and reliability. Reliability of the instrument is imperative to the consistent interpretation of the statistical tests (Field, 2013; Rovai et al., 2013). Cronbach’s alpha coefficient is effective in determining the internal consistency and the acceptable

coefficient is .70 or higher (Pallant, 2013). I examined the Cronbach’s alpha values from a pilot study described in the data collection technique section to test the reliability of the instrument. Cronbach’s alpha is a measure of internal consistency reliability based on the value of a correlation between items of an instrument (Rovai et al., 2013). Lee and Yu (2011) calculated Cronbach’s alpha value to validate the inter-item reliability of the variables related to knowledge sharing. As shown in the Table 3, the Cronbach’s alpha

coefficient of question set for employees’ willingness for knowledge sharing was .870, employees’ perception of trust was .891, organizational fairness was .884, and

supervisor’s competency was .942. The Cronbach’s alpha coefficient for each question sets of the survey exceeded the acceptable value of .700, indicating a reliable consistency.

Data Collection Technique

After IRB approval, I conducted a pilot study to examine the assumptions and the consistency of the instrument. After the assumptions and validation of the instrument were satisfied, I proceeded with the final study. The HUD public website at

http://www.hud.gov served as the source to retrieve the names and contact information of

http://www.hud.gov served as the source to retrieve the names and contact information of